Madhuri Ekbote Sheshagiri
• 8 Months of experience as data scientist.
• 1.8 years of experience as Java developer.
• As a Data Scientist analysed data sets and helped companies make decisions based on findings.
• Served as a thoughtful advisor, working with sales and team members.
• Utilized algorithmic and programming tools to build helpful predictive models.
• Performed exploratory data analysis and discovered notable relationships.
• Worked on projects using algorithms such as SVC, RANDOM FOREST, XGBOOST, LINEAR REGRESSION, LOGISTIC REGRESSION, ARIMA, SARIMA, NAIVE BAYES, K-MEANS, KNN, SVM, PCA, ARMA, ANN, DECISION TREE, K-Means Elbow Technique.
• Utilized the java programming tools and skills in developing the applications with Customer business requirements.
• Adept at data management activities, including collection, cleansing, standardization. Good Statistical Knowledge.
• Java, SQL, Python
• Jupyter, Tableau, Eclipse, JDBC, Pycharm, Matplotlib, Microsoft Excel, scikit learn, Pandas, Numpy, Seaborn. Statistics
• T-test, one-way ANOVA test, correlation matrix, ROC curve, chi square test. Certifications
• JAVA SE6 from Oracle
• Data Science from IABAC
Data Scientist at Rubixe Bangalore India
May 2019 – Present
Rubixe is a global technology company specializing in disruptive technologies – Artificial Intelligence (AI), Machine Learning, Robotic Process Automation (RPA), Block Chain and Internet of Things (IoT).
Project 1: Telecom Churn Rate Calculation
• Predicting churn rate of the customers.
• Worked with team in defining clear business problem and in understanding the variables that influencing the customers to migrate to competitors.
• Modelling ml algorithm random forest and XGBOOST, model evaluation using Gridsearch . Project 2: ITSM Incident Management
• Predict volume of incident tickets, forecasting the volume of high priority tickets.
• Involved in Exploratory Data Analysis process, worked with Pandas, Numpy, Matplotlib and seaborn In analysing the pattern present in the data.
• There was a challenges with data, as there were significant wrong entries and human English
biases were present (about 30%).
• Used Random forest classifier to create classifier with 98% accuracy.
• Applied Augmented Dickyfuller Test to check the stationarity of data and forecasted the volume of high priority tickets using SARIMA since there was a seasonality present in a data. Project 3: Forecasting Spare Parts Inventory
• The business case is on the inventory management. Keeping inventory of spare parts in various service centres to the market demand.
• Applied statistical techniques to transform the data, applied ARIMA to forecast volume of top 10 spare parts which are highly in demand with more accuracy.
Project 4: Predicting Sales effectiveness based on lead category
• Exploring Machine Learning to pre-categorize the lead quality and as a result, expecting significant increase in sales.
• Data cleaning process and understanding the data consumed 65% of our time and depending on the observation we created a new feature which categorises the various leads to high and low potential which is the target variable.
• Here we used LASSO model to check the feature importance and trained the model using RandomForest Classifier with an accuracy of 73%.
Hackathon Projects on Data Science:
• House Price Prediction
• Book Price prediction
• Flight Price Prediction
• News Categorisation
Software Engineer at IndGlobal Digital Pvt Ltd -Bangalore JUN-2012 TO AUG-2014
Indglobal is an established IT service provider in the field of Mobile applications, CRM Development, Digital Marketing, Ecommerce, Custom Web portal since 10 years expanding the branches across 4 countries.
Project 4: Bank Transaction Application
• The business case is on the bank transaction management. Keeping track of the transactions of users.
• Used my knowledge with JAVA servlets, JSP to design an application on account open, account close, Transaction, Deposit, Withdrawal for easy interaction of user with bank.
• Connecting the JDBC with SQL server in backend to perform all the operations and update in a SQL database successfully.
Project 5: Hostel Management Application
• The business case is on the hostel management. Keeping track of the hostel requirements and availability.
• Used JAVA servlets and JSP to design an application which is useful in maintaining a record of needs and availability of rooms in a hostel and also Keeping track of the student’s data in a SQL database at the backend.
• There was an issue where client requirements changed when 70% of the work was done and we can able to meet the requirement and Project was successfully delivered and deployed at client location.
Bachelors in Information Science Engineering, Rao Bahadur Y. Mahabaleswarappa College (Visvesvaraya Technological University) Bellary, India
August 2008 – April 2012